Daniel W Boston1, Jarom E Sauble. 1. Department of Restorative Dentistry, Temple University School of Dentistry, 3223 North Broad Street, Philadelphia, PA 19140, USA. daniel.boston@temple.edu
Abstract
PURPOSE: To assess the performance of laser fluorescence (DIAGNOdent, "DD") for differentiating carious versus non-carious dentin during caries excavation using caries dye as the gold standard. METHODS: The superficial carious dentin surfaces of lesions in 25 teeth were exposed, and the surface evaluated by DD. Next, caries dye was applied and the surface was observed for dye staining. The lesion was then excavated approximately 0.5 mm with a round bur. The resulting surface was evaluated by DD, and the above sequence was repeated until three non-stained surfaces were evaluated by DD. Using "caries dye staining" as the gold standard, correct classification proportion was calculated at each DD cutpoint in the dataset. Using the optimal cutpoint, sensitivity, specificity and likelihood ratios were calculated. Non-parametric ROC analysis was performed and AUC calculated. RESULTS: With an optimal cutpoint of 11 or 12, DD was highly accurate in discriminating between dye-stainable and dye-unstainable surfaces, producing a correct classification proportion of over 98% and an AUC of 0.999 for all 176 surfaces. These results support further refinement of DD for use in caries excavation and further research relating DD to other caries excavation endpoints.
PURPOSE: To assess the performance of laser fluorescence (DIAGNOdent, "DD") for differentiating carious versus non-carious dentin during caries excavation using caries dye as the gold standard. METHODS: The superficial carious dentin surfaces of lesions in 25 teeth were exposed, and the surface evaluated by DD. Next, caries dye was applied and the surface was observed for dye staining. The lesion was then excavated approximately 0.5 mm with a round bur. The resulting surface was evaluated by DD, and the above sequence was repeated until three non-stained surfaces were evaluated by DD. Using "caries dye staining" as the gold standard, correct classification proportion was calculated at each DD cutpoint in the dataset. Using the optimal cutpoint, sensitivity, specificity and likelihood ratios were calculated. Non-parametric ROC analysis was performed and AUC calculated. RESULTS: With an optimal cutpoint of 11 or 12, DD was highly accurate in discriminating between dye-stainable and dye-unstainable surfaces, producing a correct classification proportion of over 98% and an AUC of 0.999 for all 176 surfaces. These results support further refinement of DD for use in caries excavation and further research relating DD to other caries excavation endpoints.